Troubleshoot Poor RF Issues Quickly with Intel® Connectivity Analytics

 Intel’s corporate wireless local area network (WLAN) is crucial to employee productivity. If users experience poor radio frequency (RF) conditions, productivity can plummet while user frustration escalates. We have found that reliable WLAN service starts with good access point (AP) deployment—which, in turn, is the result of effective planning. But as you can read in our recent white paper, “Optimizing and Troubleshooting Wi-Fi Networks Using Client Analytics,” planning alone isn’t sufficient to achieve the network reliability we need.

Intel IT’s Best Practices for Avoiding Poor RF Issues

Intel IT has developed a few guidelines for ideal AP placement:

• Place an AP every 45 feet.
• Place an AP near all four corners of small conference rooms.
• For larger conference rooms, install APs inside the conference room, one per every eight users.

We estimate that these guidelines prevent poor RF coverage about 80% of the time, but they don’t always meet our goal of 99.999% coverage. It’s difficult to simulate internal walls, which can vary from cubicle panels to concrete. Also, APs use a dynamic power level that can change depending on the users’ physical distribution around each AP. Another problem is that room layouts might change between planning AP placement and installation. Therefore, after a third-party firm installs APs based on the above guidelines, we go the extra mile to validate that the installed APs provide the necessary coverage.

Traditional RF Coverage Validation Is Time-consuming and Costly

RF coverage validation is traditionally performed in one of two ways:

• Use a commercial software application to simulate coverage, which may not reflect the floor layout accurately unless we manually update it.
• Physically walk the entire floor with a client device. However, this is time-consuming and not all campuses have on-site IT staff. 

Not satisfied with either of these two options, a few years ago we developed an in-house alternative to walking the floor. Our customized software displays AP power levels and user counts per AP in a dashboard (see screenshot below). We can look for signs of low coverage—APs with a high power level and a high user count. In the following screenshot, a glance down the rows of data reveals that the AP named etb671y-aej76 (ninth row) meets these criteria. 

 However, our in-house software wasn’t completely ideal. Every five minutes, we had to manually collect the data from the WLAN controller across the entire company. We also needed to develop sophisticated algorithms to eliminate non-prime-time data, define interesting values, and accommodate geographic regions’ differences (for example, in some regions, Friday is a working day; in others, it is not). With hundreds of thousands of APs installed worldwide, such manual efforts are nearly impossible. 

This approach to network validation does not scale well and uses precious IT resources. 

A Faster, Better Approach to Validation

Intel® Connectivity Analytics, recently developed in collaboration with Cisco and Intel’s Client Computing group, is a cost-effective, highly scalable alternative approach to network validation.

Prior to using Intel Connectivity Analytics, we could gather only AP data—there was no visibility into the client side of things. AP data could tell us about the number of users and transmit (Tx) power level, but we merely used this information as markers for low coverage. It wasn’t possible to measure the coverage directly from the client perspective. 

In contrast, Intel Connectivity Analytics constantly gathers data from client devices themselves. The client data represents the actual client experience.

This solution provides granular driver-level wireless client insights for any client using Intel® Wi-Fi products while connected to a supported Cisco wireless network. The APs automatically collect the Wi-Fi telemetry sent from the clients and forward it securely to a Cisco Catalyst controller, Cisco Catalyst Center management platform, or a Cisco Meraki Cloud management platform. The solution then processes the data and presents network managers with intelligent reports and insights.

Intel Connectivity Analytics provides an efficient, targeted approach that identifies low received signal strength indicator (RSSI) events (see screenshot below) that pinpoint users who are experiencing poor RF. We can explore the issue directly by examining the specific AP power level history and deployment without needing to collect and analyze vast amounts of data.

By drilling down, we see that the same AP is the source of all the low RSSI events on a particular date (see screenshot below). Note: RSSI is measured in decibel milliwatts (dBm), where a larger negative number represents a lower signal strength. For example, an RSSI of -70 dBm is stronger than an RSSI of -85 dBm. An RSSI of -50 dBm is a strong signal, while -85 is very weak. 

 Further drill-down shows that this AP is transmitting at the highest power level, suggesting insufficient coverage in the specific area.

Finally, floor map analysis discovered a lack of AP neighbors close to this AP, resulting in poor signal strength for clients.

 Conclusion

Intel Connectivity Analytics lets us quickly find low-coverage areas and pinpoint the cause. It eliminates the need to walk the floor or manually collect and analyze RF data from the entire company. We can now focus our efforts only where it matters.

The time savings are enormous. On the client side, troubleshooting a connectivity issue without Intel Connectivity Analytics could take 10–15 minutes, and the IT administrator could fix only one client at a time. Intel Connectivity Analytics has helped mean time to repair drop significantly to 10–15 seconds, and the solution can simultaneously identify numerous problematic connections. For network issues, it might take days to diagnose what is causing a rash of connectivity problems. With Intel Connectivity Analytics, we can now discover the root cause of a problem in a matter of seconds. In addition, we no longer need to support our in-house tool or create internal custom scripts and automation jobs for collecting and managing client data.

For information about how Intel IT uses Intel Connectivity Analytics to improve the Wi-Fi network user experience and network management efficiency, read the IT@Intel white paper, “Optimizing and Troubleshooting Wi-Fi Networks Using Client Analytics.”